Biodiversity is a key component of agricultural ecosystems and belongs to one of the exceeded planetary boundaries. In the quest for innovation to mitigate impact, it is essential to have tools to assess the anthropogenic impact on biodiversity. In this study, we developed a new indicator, I-BIO that aimed to predict the impacts of management practices and the influence of the landscape on overall biodiversity at species group level. I-BIO encompassed four groups: microorganisms, vegetation, invertebrates and vertebrates. This predictive indicator was built using DEXi software that facilitates the design of hierarchical decision trees, based on input variables in the form of qualitative classes, aggregated using utility functions. I-BIO was constructed using data from a longterm platform in Scotland and then applied to two data sets in France and one in Scotland. The results showed that the lowest class is predominant for all branches especially for invertebrates and soil invertebrates. Monte Carlo sampling was used to test for sensitivity. Then, a validation step was carried out to compare I-BIO outputs with real data using data on the four groups from Scottish and French datasets. I-BIO predictions of weed and soil invertebrate abundance conformed to field data. Due to low numbers of samples, validation of microorganisms, flying invertebrates and vertebrates predictions require more data. Future developments of I-BIO will focus on predictions specifically to assess the impact of management on species richness. Avenues for improvement include a better integration of landscape features and the history of agricultural practices as they both have a great influence on biodiversity.